Discussion
In this study we tested for signatures of post-mating selection on gamete traits in relation to population density, and possible interactive effects of population density and sperm concentration on sperm motility and fertilisation rates. Our study yielded four key results: 1) males from high density populations produce smaller sperm compared with males from low density populations, although we found no difference in egg size between females from different population densities; 2) females from low density populations have lower fertilisation success, although this becomes less important as sperm concentration increases; 3) variances in fertilisation success were higher for females than males; and 4) gamete compatibility between males and females increases as sperm concentrations increase.
We found that males from high density populations, where sperm competition would be expected to be more prevalent (Evans and Lymbery 2020), produced smaller sperm compared to males from low density populations. To the extent that sperm production trades-off against sperm size (Parker 1982; Gomendio et al. 1991; Tourmente et al. 2011), this result supports the prediction that an increase in the overall magnitude of sperm competition should select for a greater number of smaller sperm, while lower sperm competition risk should result in the production of fewer longer sperm (Parker 1982; Stockley et al. 1997; Gage and Morrow 2003; García-González and Simmons 2007; Lüpold et al. 2020). These results support comparative studies of fish that show sperm length decreased with sperm competition risk (Stockley et al. 1997), although some studies have found the opposite, with among species comparisons showing a positive relationship between sperm competition risk and sperm size (Gage 1994; Byrne et al. 2003; Lifjeld et al. 2010; Johnson et al. 2013). The reason for such varying and contrasting results likely arises from the complex relationship between sperm morphology and sperm performance (e.g. swimming speed and/or fertilisation potential). In some species (especially internally fertilisation taxa) longer sperm may be associated with improved swimming performance and provide a competitive advantage under sperm competition (Gomendio et al. 1991; Parker et al. 2010; Johnson et al. 2013). However, sperm length may also be associated with sperm longevity, with longer, larger sperm living longer than smaller sperm. If so, producing longer-lived sperm may confer an important advantage in broadcast spawners in low density populations, allowing more time for sperm-egg encounters and increased fertilisation success (however see, Levitan 2000). A study of the broadcast spawning tubewormGaleolaria gemineoa showed that sperm with small heads but long tails were favoured in high-concentration environments, whereas sperm with long heads were favoured at low concentrations and old ages (Johnson et al. 2013). The longer sperm produced by males from low density populations detected in this study may provide an advantage if they have greater longevity, although longevity was not directly assessed in this study. Our analysis also showed that longer sperm from low density populations did not show greater motility or swimming speeds, suggesting that the larger size did not confer a swimming advantage. The relation between sperm size and fertilisation success under sperm competition is complex and appears to be highly context dependent and likely to covary with sperm swimming traits and longevity.
Theoretical and empirical studies predict that egg size should reflect adult density, with females from low-density populations producing larger eggs than their high-density counterparts, because larger eggs present larger targets for searching sperm (Levitan 2006; Crean and Marshall 2008; Evans and Lymbery 2020). In the present study we found no difference in mean egg size between females from high- and low-density populations. However, egg size is not the only mechanisms available to females for increasing the effective target size of their eggs. For example, it is well established that the eggs of many broadcast spawners release sperm chemoattractants, which are thought to increase the effective target size of eggs, thus making them more ‘visible’ to searching sperm (e.g. for theory see Jantzen 2001). While we currently lack explicit evidence that females can facultatively adjust the production of sperm chemoattractants to match the fertilisation environment, or that selection may favour increased production of such attractants in populations with persistently low adult densities (and therefore sperm limited), the idea of testing for differences either in the composition or volume of sperm chemoattractants across populations has merit. Such an adaptation may represent a more cost-effective mechanism for increasing the target size of eggs compared to increasing the structural size of eggs in sperm-limited environments and offers an exciting area for future research.
We assessed fertilisation rates of females from low- and high-density populations across a range of sperm concentrations. As population densities decrease and the risk of sperm limitation increases, we expected that females would produce eggs that are more readily fertilised (i.e. greater fertilisation success at lower sperm concentrations), while in high sperm density environments females should increase ovum defences to reduce the risk of polyspermy (Frank 2000; Firman and Simmons 2013; Kosman and Levitan 2014). Surprisingly, we found no evidence to support these ideas. Indeed, we found that females from high density sites produced eggs that were more readily fertilised at lower sperm concentrations than eggs from females originating from low density populations, although this difference became progressively less apparent as sperm concentrations increased. Furthermore, our finding that at the highest sperm concentration (2 × 108 sperm ml-1), females from both high- and low-density populations experienced similar magnitudes of abnormally fertilised eggs fails to support the idea that females from high density populations are better at blocking polyspermy compared to females form low density populations (Kosman and Levitan 2014).
Interestingly, we found that the variance components for (normal) fertilization rates across the range of sperm concentrations were generally higher in females than in males, but also that variances for females were initially high at lower sperm concentrations (2 × 102 & 2 × 103ml-1), decreased when sperm concentrations were at an intermediate level (2 × 104 & 2 × 105 ml-1), before rising sharply when sperm concentrations exceeded 2 × 106ml-1. Qualitatively similar patterns have been reported in other broadcast spawning invertebrates (Levitan 2004), suggesting that under sperm limitation, where average fertilisation rates are low, there will be heightened opportunity for selection on female traits that improve fertilisation rates. Under such conditions, for example, we might expect stronger selection for increased egg size and/or the heightened production of chemoattractants that increase sperm-egg encounter rates (see Evans and Lymbery 2020). Under intermediate sperm concentrations, fertilisation rates were generally high (~80%) and the variance in normal fertilisation rates was consequently very low. By contrast, despite high overall fertilisation rates in the highest sperm concentration groups (>2 x 106 ml-1), the proportion of normally fertilised eggs decreased dramatically at high sperm concentrations with a concomitant increase in the variance in normally fertilised eggs (see Fig. 1a). Together these findings suggest that the opportunity for selection on egg traits will depend critically on local spawning conditions, and that such patterns may be reflected on a broader spatial scale in divergent populations with varying adult densities.
Our fertilisation assays indicated that as sperm concentration increased, gametic compatibility effects became increasingly important. This suggests that under sperm limited conditions, where the risk of fertilisation failure is higher, any benefits associated with selecting genetically compatible sperm (Kosman and Levitan 2014; Oliver and Evans 2014) are offset by the direct costs of leaving many eggs unfertilised. However, as sperm concentration increases, eggs can afford to become ‘choosier’ as this will ensure that fertilisations are biased toward genetically compatible sperm (Sherman et al. 2015; Lymbery et al. 2017) whilst avoiding the direct costs associated with polyspermy. Although we have yet to understand the mechanistic basis that underlie these dynamics patterns of sperm-egg interaction, we suspect that sperm chemoattraction and gamete surface proteins play important roles in differentially regulating sperm-egg encounter rates across the sperm concentration continuum (Evans and Sherman 2013). Irrespective of the underlying mechanisms, our results provide further support to an increasing number of studies that have shown that local environmental conditions can influence the magnitude of compatibility effects between male-female combinations (Levitan and Ferrell 2006; Levitan et al. 2007; Nystrand et al. 2011; Lymbery and Evans 2013; Sherman et al. 2015; Rudin-Bitterli et al. 2018). Collectively, these studies highlight the importance of testing for compatibility effects across a range of ecologically relevant environmental conditions.
In conclusion, our study revealed significant effects of population density and sperm concentration on gamete morphology and fertilisation dynamics, respectively. While the extent of variation at the population level may be determined by both environmental and/or genetic factors, the critical importance of gamete-level natural and sexual selection in broadcast spawners (Evans and Sherman 2013; Evans and Lymbery 2020) leads us to predict that much of the variation we observe in these populations is adaptive. However, we also report highly dynamic patterns of fertilisation across experimentally altered sperm environments, highlighting the importance of phenotypic plasticity in governing sperm-egg interactions and the likely dynamic selective environment in which fertilisation plays out. We eagerly anticipate future work that seeks to understand the extent to which the mechanisms underlying these dynamic patterns of sperm-egg interaction are themselves plastic in their expression.